> ## Documentation Index
> Fetch the complete documentation index at: https://docs.agentruntime.io/llms.txt
> Use this file to discover all available pages before exploring further.

# ElevenLabs connector

> Text-to-speech, speech-to-text, voice design, music, and conversational AI agents via the ElevenLabs API in AgentRuntime workflows.

The **elevenlabs** MCP adapter connects to the [ElevenLabs API](https://elevenlabs.io/docs). Synthesize and transform audio, manage voices, compose music, and build conversational AI agents — suitable for briefings, IVR, accessibility, content repurposing, and voice-enabled agent workflows.

<Note>
  Full platform parity (27 tools) is rolling out across deployments. Some workspaces may still see a smaller tool set (`list_voices` and `text_to_speech` only) until their catalog is updated to the latest adapter version.
</Note>

## Prerequisites

* An ElevenLabs account with an **API key**
* Character / usage quota sufficient for your volume (TTS, STT, and agent tools consume credits)
* **project\_contributor** access in AgentRuntime

## Connect in AgentRuntime

<Steps>
  <Step title="Get an ElevenLabs API key">
    Sign in to [ElevenLabs](https://elevenlabs.io/) and copy your API key from profile settings.
  </Step>

  <Step title="Create a connection">
    Go to **Connections** and click **New custom connection** to create an API key connection with `APIKey`.
  </Step>

  <Step title="Add elevenlabs MCP instance">
    Go to **MCP**, click **Add instance** for **elevenlabs**, wire your connection on **Instance config**, set the profile **active**, and save.
  </Step>

  <Step title="Discover voices">
    Run **mcp\_call** → `list_voices` or `search_voices` and note a `voice_id` for TTS steps.
  </Step>
</Steps>

See [Authentication](/integrations/authentication) for binding connections to MCP instances.

## What you can build

* **Daily audio brief** — LLM summarizes news → `text_to_speech` → store or attach MP3.
* **Multilingual outreach** — Translate → `text_to_speech` with `eleven_multilingual_v2`.
* **Meeting notes** — `speech_to_text` on recorded audio with optional diarization.
* **ConvAI agent ops** — `create_agent`, `simulate_conversation`, `make_outbound_call` for voice agent workflows.
* **Music beds** — `compose_music` or `video_to_music` for short-form video scores.

## Tools

### Speech and audio

| Tool                    | Description                         |
| ----------------------- | ----------------------------------- |
| `text_to_speech`        | Synthesize speech from text         |
| `speech_to_text`        | Transcribe audio files              |
| `speech_to_speech`      | Transform audio to another voice    |
| `text_to_sound_effects` | Generate sound effects from text    |
| `isolate_audio`         | Isolate voice from background noise |

### Voices

| Tool                        | Description                                |
| --------------------------- | ------------------------------------------ |
| `list_voices`               | List voices for the API key                |
| `search_voices`             | Search the user's voice library            |
| `get_voice`                 | Get details for one voice                  |
| `voice_clone`               | Instant voice clone from audio files       |
| `text_to_voice`             | Voice design previews from a description   |
| `create_voice_from_preview` | Save a generated preview to the library    |
| `search_voice_library`      | Search the shared ElevenLabs voice library |

### Models and account

| Tool                 | Description                        |
| -------------------- | ---------------------------------- |
| `list_models`        | List available TTS/STT models      |
| `check_subscription` | Subscription status and usage JSON |

### Conversational AI (ConvAI)

| Tool                          | Description                                  |
| ----------------------------- | -------------------------------------------- |
| `create_agent`                | Create a conversational AI agent             |
| `add_knowledge_base_to_agent` | Add URL, file, or text knowledge to an agent |
| `list_agents`                 | List agents                                  |
| `get_agent`                   | Agent details                                |
| `get_conversation`            | Conversation transcript and metadata         |
| `list_conversations`          | Filtered conversation history                |
| `simulate_conversation`       | Test an agent with a simulated user          |
| `make_outbound_call`          | Outbound call via Twilio or SIP trunk        |
| `list_phone_numbers`          | Phone numbers on the account                 |

### Music

| Tool                          | Description                                       |
| ----------------------------- | ------------------------------------------------- |
| `compose_music`               | Generate music from prompt or composition plan    |
| `create_composition_plan`     | Plan-only (no generation credits)                 |
| `video_to_music`              | Score from one or more video files                |
| `upload_music_for_inpainting` | Upload audio for music\_v2 inpainting (`song_id`) |

## Example

Convert an LLM summary to audio:

```json theme={null}
{
  "id": "elevenlabs-tts",
  "type": "mcp_call",
  "name": "Synthesize briefing",
  "tool_name": "text_to_speech",
  "tool_args": {
    "voice_id": "{{input.voice_id}}",
    "text": "{{steps.llm-summary.result.text}}",
    "model_id": "eleven_multilingual_v2"
  },
  "depends_on": ["llm-summary"],
  "timeout_s": 60
}
```

## Configuration

| Key                | Required | Description                                                       |
| ------------------ | -------- | ----------------------------------------------------------------- |
| `api_key`          | Yes      | ElevenLabs API key                                                |
| `output_mode`      | No       | `files`, `resources` (default), or `both` for binary tool outputs |
| `default_voice_id` | No       | Default voice when TTS omits `voice_id`                           |
| `api_residency`    | No       | `us`, `eu`, `in`, `sg`, or `global`                               |
| `base_path`        | No       | Base directory when `output_mode=files`                           |

Connections typically map `APIKey` → `api_key` on the MCP instance.

## Troubleshooting

| Issue                         | Fix                                                                                 |
| ----------------------------- | ----------------------------------------------------------------------------------- |
| Voice not found               | Run `list_voices` or `search_voices`; IDs differ from display names                 |
| Quota exceeded                | Check usage in ElevenLabs dashboard; shorten text or upgrade plan                   |
| Garbled or clipped audio      | Split long text; respect model limits                                               |
| Wrong language                | Use `eleven_multilingual_v2` or a voice verified for the target language            |
| File path errors on STT/music | Tools that read local files need absolute paths accessible to the connector runtime |

## Related

* [Integrations quickstart](/integrations/quickstart)
* [Vapi connector](/connectors/vapi) — voice assistants and telephony
* [Connector catalog](/integrations/connector-catalog)
* [Troubleshooting](/platform/troubleshooting)